Factorial validity and invariance testing of the exercise dependence scale-revised in Swedish and Portuguese exercisers

The present study investigated the factorial validity and factorial invariance of the 21-item Exercise Dependence Scale-Revised using 162 Swedish and 269 Portuguese exercisers. In addition, the prevalence of exercise dependence symptoms and links to exercise behavior, gender, and age in the two samples was also studied. Confirmatory factor analyses demonstrated that the hypothesized 7-factor model made a good fit to data in both samples. Multi-sample analyses supported partial measurement invariance across the samples; only factor loadings involving items 3 and 19 were noninvariant. The "Reduction in Other Activities" and "Lack of Control" factors were the most problematic scales in terms of average variance explained and reliability measures of weighted omega and Cronbach's alpha. In total, 9.2% of the Swedish sample and 5.2% of the Portuguese sample were classified as at risk for exercise dependence. Overall, the results support the factorial validity of the translated Exercise Dependence Scale-Revised in samples outside North-America, although more research is needed.